updates

precompute_caches.py

@@ -28,9 +28,9 @@ def precompute_all_caches(
     
     if models_to_test is None:
         models_to_test = [
-            'gabrielloiseau/LUAR-MUD-sentence-transformers',
-            'gabrielloiseau/LUAR-CRUD-sentence-transformers', 
-            'miladalsh/light-luar',
+            # 'gabrielloiseau/LUAR-MUD-sentence-transformers',
+            # 'gabrielloiseau/LUAR-CRUD-sentence-transformers', 
+            # 'miladalsh/light-luar',
             'AnnaWegmann/Style-Embedding'
         ]
     
@@ -194,7 +194,7 @@ from utils.visualizations import visualize_clusters_plotly
 
 if __name__ == "__main__":
     # Test with a small subset first
-    instances=[i for i in range(20)]  # First 20 instances for testing
+    instances=[i for i in range(10)]  # First 20 instances for testing
     cache_stats = precompute_all_caches(
         models_to_test=[
             'gabrielloiseau/LUAR-MUD-sentence-transformers'

utils/interp_space_utils.py

@@ -544,7 +544,7 @@ def compute_clusters_style_representation_3(
     background_corpus_df: pd.DataFrame,
     cluster_ids: List[Any],
     cluster_label_clm_name: str = 'authorID',
-    max_num_feats: int = 20,
+    max_num_feats: int = 25,
     max_num_documents_per_author=1,
     max_num_authors=10,
     max_authors_for_span_extraction=4,